Rumination, but Not Reflection, Predicts the Reporting of Post-concussive Symptoms in a Non-clinical Sample

Abstract

Objective

Although traditionally associated with mild head trauma, post-concussive symptoms are commonly reported across both healthy and other clinical populations. Existing research indicates that individuals with depression report high levels of post-concussive symptoms, though the underlying causes of this association remain unknown. The current study aimed to explore potential factors underlying this relationship: specifically, how maladaptive and adaptive self-focused cognitive coping styles, namely, rumination and reflection, respectively, differentially contribute to post-concussive symptoms.

Method

489 undergraduate students and 136 community participants with no history of head trauma completed the Rivermead Post-Concussion Symptom Questionnaire, the Depression Anxiety and Stress Scales–21 Items, and the Rumination and Reflection Questionnaire.

Results

Rumination significantly predicted post-concussive symptoms after controlling for the effects of depression, demographic variables, and confounding factors. However, reflection did not predict lowered symptom reporting as hypothesized. Overall, the final model explained 42.5% of the variance in reported symptoms. Rumination, female gender, prior history of headaches, pre-existing diagnosis of attention-deficit/hyperactivity disorder, and depression were significant predictors of elevated post-concussive symptoms.

Conclusions

These findings suggest that rumination, a negative coping style linked to depression, plays a key role in influencing post-concussive symptom endorsement. Future research should examine combinations of rumination and reflection, as well as other cognitive coping styles, that affect symptom reporting. Clinically, assessment of an individual’s ruminative tendency following head trauma and the potential incorporation of rumination-focused treatment is recommended to improve recovery outcomes.

INTRODUCTION

Post-concussive symptoms (PCS) are a complex set of somatic, cognitive, and affective difficulties that can occur after a mild traumatic brain injury (mTBI) or concussion (Iverson, 2006; Pulsipher et al., 2021). Main symptoms include headaches, sensitivity to noise and light, feeling nauseous and dizzy, tiredness, disrupted sleep, irritability, emotional dysregulation, and concentration and memory complaints (Iverson & Lange, 2003; Pulsipher et al., 2021; Trahan et al., 2001). These symptoms typically appear shortly after injury and most individuals recover within a few days to 3 months (Edmed & Sullivan, 2012; Williams & Wood, 2010). A minority of individuals, however, experience persistent symptoms extending beyond 3 months post-injury, resulting in a potential diagnosis of Persistent Post-Concussive Syndrome (PPCS; Williams & Wood, 2010). Indeed, official diagnostic frameworks, including the Diagnostic and Statistical Manual Fourth Edition [DSM-IV-TR; American Psychiatric Association (APA), 2022] and the International Classification of Diseases, Tenth Edition [ICD-10; World Health Organization (WHO), 1993], have recognized PPCS as a formal diagnosis. However, the validity of the diagnosis of PPCS remains controversial, and it is no longer included the most recent version of the DSM-5-TR.

Despite its potential to significantly impair academic, occupational, and familial functioning (Iverson, 2006), the validity of PPCS as a distinct diagnostic entity remains controversial (Iverson & Lange, 2003). Such controversy stems from the non-specificity of PCS symptoms, as these symptoms are observed in individuals without a history of concussion or mTBI (Laborey et al., 2014). In fact, multiple studies have shown that PCS are as prevalent or even more prevalent, in the general population (Asken et al., 2017; Balasundaram et al., 2016; Edmed & Sullivan, 2012; Garden & Sullivan, 2010; Iverson & Lange, 2003; Wang et al., 2006). Further research has demonstrated that PCS are also reported by university athletes without a history of concussion (Asken et al., 2017; Ferguson et al., 1999) or those with long-term pain conditions (Iverson & McCracken, 1997; Smith-Seemiller et al., 2003), Post-Traumatic Stress Disorder (Foa et al., 1997), orthopedic injuries (Ettenhofer & Barry, 2012), insurance claims (Hanks et al., 2019; Tsanadis et al., 2008), or trauma unrelated to the brain (Meares et al., 2011). The lack of specificity in symptoms related to PCS underscores the importance of understanding the factors contributing to PCS symptom reporting.

Various factors have been associated with the increased reporting of PCS. These include genetic predispositions (Iverson et al., 2017), biological sex (Iverson et al., 2017), and mental health conditions (Iverson et al., 2015; Lange et al., 2010; Meares et al., 2008). Research findings consistently demonstrate that a diagnosis of depression or higher levels of depressive symptoms correlate with increased reporting of post-concussion-like symptoms (e.g., Garden & Sullivan, 2010; Iverson, 2006; Iverson & Lange, 2003; Suhr & Gunstad, 2002). At a basic level, many symptoms specific to depression mirror those captured by PCS (Trahan et al., 2001). These shared symptoms involve emotional difficulties, tiredness, being more easily frustrated and more irritable, attentional and sleeping difficulties, forgetfulness, and diminished mental efficiency (Trahan et al., 2001). A study by Iverson (2006) showed that 90% of individuals with depression reported symptoms that aligned with PCS and more than half of them satisfied DSM-IV-TR (APA, 2000) and ICD-10 (WHO, 1992) diagnostic criteria for PPCS. Thus, distinguishing whether self-reported symptoms stem from depression, persistent post-concussion syndrome, or a combination of both is exceedingly challenging because many of the symptoms overlap in these conditions (Iverson, 2006).

Beyond symptom overlap, cognitive coping styles, particularly coping mechanisms, may further explain the relationship between depression and PCS symptom reporting (Folkman et al., 1986; Lazarus & Folkman, 1984; Nagase et al., 2009). Ineffective coping strategies can heighten attention to physical and cognitive difficulties, leading to an intensified perception of disability. Depressed individuals tend to endorse a broader range of symptoms, including those inconsistent with traditional definitions of depression, such as PCS (Trahan et al., 2001). Jacobson (1995) proposed that coping processes may influence the trajectory of PCS, a hypothesis supported by studies suggesting that poor coping skills exacerbate PCS symptomatology (Soo & Sherman, 2015).

A key maladaptive coping style linked to depression is rumination (Nolen-Hoeksema, 1998), which involves persistent, repetitive, and self-focused attention on negative experiences (Watkins, 2008). Rumination is characterized by passive contemplation that perpetuates psychological distress and impairs problem-solving (Trapnell & Campbell, 1999; Watkins & Roberts, 2020). It is strongly associated with neuroticism (McCrae & Costa, 1987), a trait linked to negative emotionality and poor coping. Prospective studies have identified rumination as a significant risk factor for depression (Taku et al., 2009; Thomsen et al., 2013), it is a salient cognitive coping style of major depression (Kovács et al., 2020; Zhou et al., 2020), and rumination predicts symptom severity, duration, and the onset of major depressive episodes (Kuehner & Weber, 1999; Nolen-Hoeksema, 2000; Nolen-Hoeksema et al., 1994; Nolen-Hoeksema & Morrow, 1991). Furthermore, rumination is associated with impaired problem-solving, heightened recall of negative experiences, and exacerbated feelings of helplessness (Lyubomirsky et al., 1998; Caldwell & Nolen-Hoeksema, 1998; Lyubomirsky & Nolen-Hoeksema, 1995; Nolen-Hoeksema et al., 1999). Given its role in perpetuating depression and its link to negative cognitive appraisal, rumination may contribute to the maintenance and reporting of PCS. This relationship is particularly important in the context of concussion, such that rumination may contribute to maladaptive appraisals of recovery and amplify perceived symptoms.

In contrast, reflection represents an adaptive form of self-focused attention driven by curiosity and intellectual engagement with one’s experiences (Boyraz & Efstathiou, 2011; Silvia et al., 2005; Trapnell & Campbell, 1999). Unlike rumination, reflection facilitates active problem-solving, enhances self-awareness, and fosters psychological resilience (Takano & Tanno, 2009; Trapnell & Campbell, 1999; Watkins, 2008). Research indicates that reflection is associated with lower levels of depression (McFarland et al., 2007; Takano & Tanno, 2008; Treynor et al., 2003), improved mood (Boyraz & Efstathiou, 2011), and reduced suicidality (Crane et al., 2007). Furthermore, reflection enhances self-knowledge and psychological adjustment (Martin & Tesser, 1996; Trapnell & Campbell, 1999). Although research on protective factors in PCS is limited (Kersivien et al., 2024), one study found that self-reported trait resilience (i.e., the tendency to “bound back” from stress) significantly predicts lower PCS symptomatology, even after controlling for demographic and clinical variables such as sex and ADHD (Kersivien et al., 2024). As reflection is associated with positive self-perception and resilience (Newman & Nezlek, 2019; Trapnell & Campbell, 1999), it may safeguard against PCS symptom reporting.

Despite extensive evidence linking depression to increased PCS symptom reporting (Edmed & Sullivan, 2012; Iverson, 2006; Iverson et al., 2020; Lambert et al., 2022), the cognitive mechanisms underlying this relationship remain unclear. This study aims to investigate cognitive coping mechanisms in the depression–PCS relationship, with a specific focus on rumination and reflection. It was hypothesized that rumination, a key coping style linked with depression (Spasojević & Alloy, 2001; Nolen-Hoeksema, 1998; Kovács et al., 2020; Zhou et al., 2020), would positively predict greater reporting of PCS. As reflection has been associated with a more positive self-perception (Boyraz & Efstathiou, 2009; Newman & Nezlek, 2019), which is thought to be beneficial against PCS (Kersivien et al., 2024), it was hypothesized that individuals who engage in greater reflection would report fewer PCS. It was also hypothesized that rumination and reflection would be predictive of PCS above and beyond the effects of symptoms of depression. By examining these cognitive mechanisms, this study seeks to clarify the role of maladaptive and adaptive self-focus in PCS symptom reporting, potentially informing clinical approaches to PCS assessment and management. Although our sample comprised individuals without a history of head injury, understanding the cognitive styles that influence PCS symptom reporting in non-clinical populations provides important insights into mechanisms that may contribute to persistent symptoms following mTBI/concussion.

MATERIALS AND METHODS

Participants

Participants in this study were recruited from two sources. First, undergraduate psychology students from the University of New South Wales and Western Sydney University participated via the SONA research recruitment portal in exchange for course credit. Additionally, community members were recruited through online advertisements on social media (e.g., X and Instagram) without compensation. As this was an online cross-sectional study, there were no restrictions on sample size. Inclusion criteria required participants to be aged between 17 and 70 years, have normal or corrected vision, be proficient in spoken and written English, and have stable internet access. Exclusion criteria included being aged less than 17 years or older than 70, and a history of neurological concern (e.g., brain injury, stroke, epilepsy, tumors, or concussion). A total sample of 633 individuals completed the survey (142 men, 480 women, five non-binary individuals, five transgender individuals, and one individual identifying as “other”). However, two individuals reported their age as being less than 17 years and were subsequently removed from the dataset. The final sample included 139 community members and 494 students. The age of the entire sample was 21.14 years (SD = 5.83), with an average length of formal education of 13.39 years (SD = 1.75). Ethics approval was obtained from the Human Research Ethics Committees of the University of New South Wales and Western Sydney University. Although the current study was exploratory, a post hoc power analysis indicated sufficient power (>90%) to detect small-to-moderate effect sizes (f² = 0.02).

Measures

Rivermead Post-Concussion Symptoms Questionnaire

The 16-item Rivermead Post-Concussion Symptoms Questionnaire (RPCSQ) (King et al., 1995) was used to assess the presence of post-concussive symptoms over the previous 24 h. Its items collectively examine three domains, including physiological (e.g., vomiting), cognitive (e.g., forgetfulness), and affective (e.g., feeling frustrated) symptoms. Participants scored all items on a 5-point Likert scale from not experienced (0) to severe problem (4). An overall score is derived from summing all item scores, with higher scores reflecting more severe symptomatology. The RPCSQ has been widely used in research examining mTBI and PCS with excellent psychometric properties, including internal consistency (Plass et al., 2019), strong construct validity, and inter-test reliability (Eyres et al., 2005; King et al., 1995; Summerell et al., 2023). The internal consistency of the RPCSQ in this study was excellent (α = .93).

Depression Anxiety and Stress Scales–21 Items

Participants’ emotional states of depression, anxiety, and stress were measured using the Depression Anxiety and Stress Scales–21 Items (DASS-21; Lovibond & Lovibond, 1995). This scale comprises three subscales (depression: DASS-D; anxiety: DASS-A; stress: DASS-S), each with seven items. Participants rated the extent to which they experienced specific symptoms (e.g., “I felt I was close to panic”) over the previous week on a 4-point Likert scale ranging from 0 (“Did not apply to me at all”) to 3 (“Applied to me very much, or most of the time”). Higher scores indicate greater levels of negative affect. The DASS-21 has been widely validated, demonstrating high internal consistency and strong reliability (Bibi et al., 2020; Sinclair et al., 2012). It has also shown high construct, discriminant, and convergent validity (Coker et al., 2018; Henry & Crawford, 2005). Factor analyses have supported its three-factor structure, along with a secondary factor representing overall psychological distress (Henry & Crawford, 2005). For the purpose of this study, the subscale of depression was used. The internal consistency of the entire DASS (α = .94), and the factor of depression (α = .91), were excellent in this study.

Rumination and Reflection Questionnaire

The 24-item Rumination and Reflection Questionnaire (RRQ) (Trapnell & Campbell, 1999) measures the tendency to engage in self-focused thought, divided into two 12-item subscales: rumination (repetitive focus on distressing thoughts) and reflection (self-examination driven by curiosity). Participants rated their agreement with each item on a 5-point Likert scale from 1 (Strongly disagree) to 5 (Strongly agree). Higher scores on each subscale indicate greater levels of rumination or reflection. The RRQ has demonstrated strong psychometric properties, including excellent internal consistency and robust convergent and discriminant validity (Takano & Tanno, 2008; Trapnell & Campbell, 1999). The internal consistency of the entire RRQ (α = .87), and the two subfactors of rumination (α = .91) and reflection (α = .88), were good to excellent in this study.

Procedure

This is a cross-sectional observational study that was conducted online. Data collection for this study was conducted on an ongoing basis from September 2020 to November 2023. The online survey platform Qualtrics was used to collect the study data. An accessible survey link was either embedded in the online advertisements for community individuals or provided through the online university recruitment system (SONA) for students. At the beginning of the survey, participants were provided information about the study before providing informed consent. Participants who provided consent proceeded to complete a demographics questionnaire (e.g., age, gender, years of education), as well as questions about their medical and psychiatric history. They then completed the RPCSQ, DASS-21, and RRQ.

Data Analysis

The Statistical Package for the Social Science (SPSS) was used to conduct all analyses in the current study, with p < .05 being the significance threshold. Pearson’s correlation analysis was initially used to obtain correlations between the total RPCSQ score, the DASS-D (depression) score, and scores on the rumination and reflection subscales of the RRQ. Consistent with prior research (Kersivien et al., 2024), Pearson’s correlation analysis and independent-sample t-tests between background and outcome variables (i.e., total RPCSQ score, DASS-D score, rumination and reflection subscale scores) were conducted to identify potential confounding factors, which were subsequently accounted for in the main analyses.

Hierarchical regression was the main statistical analysis method used to investigate the effects of different predictor variables on PCS symptom reporting (as measured by the total RPCSQ score). Level one of the model included demographic variables, including gender. The second level incorporated confounding factors identified in the preliminary analyses. The DASS-D score, measuring the emotional state of depression, was included in the third level of the model to account for the contribution of depression to PCS reporting. Lastly, after controlling for the effects of all other variables, rumination and reflection subscale scores were entered into levels four and five of the analyses, respectively, to explore how these two opposing forms of self-focused attention differentially contribute to PCS. Rumination was entered into the regression model prior to reflection due to the stronger theoretical and empirical links with depressive symptomatology and PCS reporting. Reflection was treated as exploratory and therefore entered last.

RESULTS

Preliminary Analyses

Data distribution

Data distribution was examined to investigate the assumptions of a regression analysis. Linearity and homoscedasticity were satisfied by examining the residual plots of the overall fitted model and each of the continuous variables (Hills, 2011; Poole & O’Farrell, 1971). In addition, variance inflation factors (VIFs) were used to assess intercorrelations between predictors in the regression model. All VIF scores were below 5, indicating satisfactory levels of multicollinearity (Tabachnick, 2007). Lastly, visual inspection of the Q-Q plot of standardized residuals from the regression model (Poole & O’Farrell, 1971) show that the assumption of normality of residual was satisfied.

Final sample

Three individuals disclosed a prior history of brain tumor, two had a history of epilepsy, and one had a history of stroke. These individuals were removed from analyses to limit the confound of potential neurological history on affecting the results. Those who completed the entire survey in less than 5 min (n = 10) were also removed from analyses to limit the potential impact of non-credible responding. The final sample comprised 625 individuals, with 136 from the general community and 489 from the student participation pool. There were 141 males, 474 females, 4 non-binary individuals, 5 transgender participants, and 1 who identified as “other.”

Sample characteristics and differences

Table 1 presents the descriptive statistics of outcome measures and background factors, including demographic and clinical variables, for the student and community samples, respectively. Out of the final sample of 626 participants, 89% identified as right-handed (n = 556). Seventy-five-point five percent of the sample identified that English was their first language, with 21.3% of the final sample identifying that they predominately spoke another language at home (n = 98). The most common first languages spoken by the sample were Chinese/Mandarin/Cantonese (n = 59), Arabic (n = 31), Hindi (n = 13), and Vietnamese (n = 11).

Table 1.

Comparison of student (n = 498) and community (n = 136) groups on background factors and outcome measures

	Student (n = 489)	Community (n = 136)	Sig.	Total (n = 625)
	Mean (SD)	Mean (SD)	p-value	Mean (SD)
Demographic
Age (years)	20.56 (4.47)	22.75 (7.41)	<.001	21.00 (4.98)
Education (years)	13.48 (1.64)	13.05 (2.00)	.024	13.40 (1.73)
Gender (n)	M = 113 F = 368, B = 3 T = 5	M = 27, F = 104 B = 1 O = 1	.195	M = 141 F = 474, B = 4 T = 5 O = 1
Clinical (n)
ADHD	20	6	.511	Y = 26
Headaches	Y = 77	Y = 32	.063	Y = 108
Outcome
PCS	12.96 (11.79)	14.46 (12.13)	.195	12.60 (10.90)
Depression	10.20 (10.10)	9.62 (9.18)	.552	4.61 (4.42)
Rumination	42.70 (10.05)	41.28 (10.64)	.160	42.30 (10.10)
Reflection	41.48 (8.80)	41.28 (8.98)	.359	41.80 (8.89)

Note: SD = standard deviation; n = number; Sig = significance; M = male; F = female; B = non-binary; O = other; T = transgender, ADHD = attention-deficit/hyperactivity disorder.

Community participants were significantly older than students, F(1, 145) = 9.81, p < .005 d = 0.42, although students reported significantly greater years of education than the community participants, F(1, 173) = 6.93, p < .01, d = 0.25. No significant group differences emerged across the remaining variables (see Table 1). Thus, there were no significant differences between the student and community samples with respect to the main outcome variables (i.e., PCS, depression, rumination, and reflection). As such, the recruitment source was not controlled for in subsequent analyses.

Correlation analyses

Depression was positively correlated with PCS such that greater depression was associated with increased PCS symptom reporting (p < .001). Rumination was positively correlated with PCS (p < .001) and depression (p < .001). Reflection was not significantly correlated with PCS or depressive symptoms. Lastly, years of education were positively correlated with age (p < .001).

Assessment of confounds

Females endorsed higher levels of post-concussive symptoms [t(309.688) = −682, p < .005, d = −0.55], together with greater symptoms of depression [t(281.55) = −3.38, p < .0005, d = −0.33] and rumination [t(602) = −4.23, p < .0005, d = −0.42]. Males and females did not differ in their degree of reflection [t(598) = −608, p = .543, d = −0.059]. Those with a history of ADHD [t(21.2) = −3.27, p = .004, d = −0.77] or headaches [t(139.94) = −6.07, p < .0005, d = −0.72] endorsed higher levels of post-concussive symptoms. Meanwhile, individuals with a history of headaches reported higher depression scores [t(149.34) = −2.76, p < .005, d = −0.30] and higher levels of rumination [t(156.22) = −3.55, p < .0005, d = −0.38]. Given these significant findings, the subsequent regression analyses included gender, and a history of ADHD or headaches, in order to adjust for potential confounding factors.

Table 2.

Pearson’s correlations of main outcome variables and demographic factors

	PCS	Depression	Rumination	Reflection	Age	Edu
PCS	—
Depression	0.608***	—
Rumination	0.341***	0.427***	—
Reflection	−0.043	−0.078	0.079	—
Age (years)	0.019	−0.002	−0.051	0.008	—
Edu (years)	−0.021	−0.055	0.035	0.058	0.248***	—

Note: PCS = post-concussive symptom; Edu = education ^***p < .001.

Hierarchical Regression Analyses

As shown in Table 3, hierarchical linear regression analyses were conducted using the total RPCSQ score as the dependent variable, with predictors separately entered at four steps. The first model included gender. This model significantly predicted the reporting of PCS and accounted for 4.4% of the variance in the total RPCSQ score, R² = .044, F(1, 595) = 27.19, p < .0005. In this model, gender emerged as a significant individual predictor of PCS (β = .21, t = 5.21, p < .0005), such that the female gender was a positive predictor of PCS reporting.

Table 3.

Hierarchical regression models predicting the postconcussive symptoms from demographic (Model 1), clinical factors (Model 2), depression (Model 3), rumination (Model 4), and reflection (Model 5) (n = 625)

	Model 1		Model 2		Model 3		Model 4		Model 5
R 2	.044***		.124***		.421***		.425***		.426***
ΔR2	.044***		.80***		.297***		.004*		.001
Predictors	β (SE)	B	β (SE)	B	β (SE)	B	β (SE)	B	β (SE)	B
Constant	4.08 (1.81)		3.48 (1.74)		0.65 (1.42)		−2.25 (1.98)		−1.47 (2.54)
Gender	5.04 (0.97)	0.21***	4.45 (0.93)	0.19***	2.61 (0.76)	0.11***	2.43 (0.77)	0.10***	2.43 (0.77)	0.101***
ADHD			8.04 (2.31)	0.13***	3.63 (1.9)	0.06	3.79 (1.89)	0.06*	3.80 (1.89)	0.06*
Headaches			7.56 (1.19)	0.25***	5.79 (0.97)	0.19***	5.58 (0.97)	0.18***	5.59 (0.97)	0.18***
Depression					0.67 (0.034)	0.56***	0.63 (0.04)	0.53***	0.63 (0.04)	0.53***
Rumination							0.09 (0.04)	0.07*	0.09 (0.04)	0.08*
Reflection									−0.02 (0.04)	−0.02

Note: ADHD = attention-deficit hyperactivity disorder; SE = standard error. ^*  p < .05 ^**p < .005 ^***  p < .0005.

The second model included a prior diagnosis of ADHD or a history of headaches. Their inclusion explained an additional 8.0% of variance in the model (ΔR² = .08, p < .0005), and the new model accounted for 12.4% of the total variance in the reporting of PCS [F(3, 595) = 27.99, p < .0005]. On an individual level, ADHD (β = .13, t = 3.49, p < .001) and a history of headaches (β = .25, t = 6.37, p < .001) significantly and positively predicted PCS, in addition to gender (β = .19, t = 4.78, p < .001).

Depression (i.e., DASS-D) was entered into level three to account for its unique contribution to the reporting of post-concussive symptoms. This addition significantly increased the predictive power of the model, accounting for an additional 29.7% of the total variance in PCS (ΔR² = .297, p < .0005). Overall, the model explained 42.1% of the variance in PCS, R² = .421, F(4, 595) = 107.45, p < .0005. Variables in this model that were significant and positive predictors of PCS symptom included depression (β = .56, t = 17.41, p < .0005), a history of headaches (β = .19, t = 5.97, p < .0005), and gender (β = .11, t = 3.42, p < .0005). ADHD was no longer a significant predictor once these other factors were included in the model (β = .06, t = 1.91, p = .056).

Rumination and reflection were entered at the final steps, completing the model with all previously entered predictor variables. As noted earlier, rumination was entered in Step 4, prior to reflection, based on its stronger theoretical and empirical associations with depression (e.g., Nolen-Hoeksema, 1998; Watkins & Roberts, 2020) and with post-concussive symptoms. In the current sample, rumination was significantly correlated with both PCS and depressive symptoms, whereas reflection was not significantly correlated with either. Given this pattern and its more exploratory nature in the present study, reflection was entered last (Step 5) to examine whether it contributed additional variance beyond established predictors. The addition of rumination significantly increased the predictive power of the model and accounted for an additional 0.4% of the total variance in PCS (ΔR² = .04, p < .05). Overall, the model explained 42.5% of the variance in PCS, R² = .425, F(5, 595) = 87.33, p < .0005. The addition of reflection did not significantly increase the predictive power of the model (ΔR² = .00, p = .63) despite the overall model remaining significant R² = .426, F(6, 595) = 72.72, p < .0005. In the final model, on an individual level, rumination (β = 0.08, t = 2.14, p < .05), depression (β = 0.53, t = 14.93, p < .0005), ADHD (β = 0.06, t = 2.01, p < .05), a history of headaches (β = 0.18, t = 5.74, p < .0005), and gender (β = 0.10, t = 3.17, p < .01) were all significant positive predictors of PCS. However, reflection (β = −0.02, t = −0.49, p = .625) was a non-significant predictor of PCS. To ensure that the order of entry for the variables did not obscure any potential effects of reflection, the analysis was also conducted with reflection entered before rumination. This alternative model yielded an identical pattern of finding, such that reflection remained non-significant, and rumination remained a significant predictor of PCS above and beyond the effects of demographic, confounds, and symptoms of depression.

DISCUSSION

Prior research has established a strong association between depression and post-concussive symptoms (Lambert et al., 2022; Thomas et al., 2022). The present study investigated whether rumination and reflection, two cognitive coping styles linked to depression, uniquely predict PCS symptom reporting. Overall, higher levels of rumination significantly predicted greater PCS, even after controlling for background factors (i.e., gender, ADHD, history of headaches) and current depressive symptoms. However, reflection did not significantly predict lower PCS symptom reporting, and as such the second hypothesis was not supported. These findings provide novel evidence that maladaptive cognitive styles, particularly rumination, may contribute to heightened PCS endorsement, independent of shared depressive symptoms. Additionally, several demographic and clinical variables, such as female gender, a prior ADHD diagnosis, a history of headaches, and current depression symptoms, were significant predictors of PCS symptom severity. These results align with prior research showing that certain individual differences and pre-existing conditions heighten PCS vulnerability (Iverson et al., 2015; Iverson et al., 2017; Kersivien et al., 2024; Thomas et al., 2022).

As hypothesized, greater rumination predicted increased reporting of post-concussive symptoms. This finding provides the first experimental evidence demonstrating the crucial role of rumination, a negative coping style linked to depression (Zhou et al., 2020; Kovács et al., 2020; Nolen-Hoeksema, 1998; Nolen-Hoeksema et al., 1994), in influencing PCS symptom endorsement, reinforcing the idea that negative, repetitive thought processes amplify symptom perception. This finding supports prior research suggesting that rumination sustains psychological distress by focusing attention on one’s difficulties rather than engaging in active problem-solving (Erickson, Newman & Tingey, 2020). Indeed, existing research has shown that overlapping symptoms between depression and PCS (e.g., fatigue, irritability, concentration difficulties) are unlikely to explain this association (Trahan et al., 2001). To this end, the present study supports previous suggestions that poor coping styles may be a mechanism through which individuals with mental health conditions report higher levels of PCS (Iverson et al., 2020).

Several psychological mechanisms may explain why rumination predicts elevated PCS symptom reporting. A potential explanation for this relationship is somatic symptom amplification, wherein individuals with heightened negative emotional states, such as depression, experience greater sensitivity to bodily sensations and misattribute normal physical discomfort to illness (Barsky et al., 1991). Rumination may intensify this process, leading individuals to focus excessively on minor discomforts, headaches, or cognitive lapses, which in turn increases PCS symptom reporting. This aligns with studies showing that negative attentional biases toward bodily sensations are a core feature of both rumination and somatization (Kang et al., 2025). Additionally, catastrophizing and health anxiety may also play a role. Catastrophizing is a cognitive distortion in which individuals exaggerate the negative appraisals of a range of events or states—including an exaggerated view on the severity of symptoms—while health anxiety leads to heightened vigilance and concern about one’s health status (Donnell et al., 2012). Given that rumination has been linked to excessive worry (Segerstrom, Tsao, Alden & Craske, 2000) and dysfunctional illness beliefs (de Jong-Meyer, Beck & Riede, 2009), it is possible that individuals who ruminate frequently also engage in catastrophizing and health-related anxiety, further reinforcing PCS symptom endorsement. Future research should examine whether these factors mediate the relationship between rumination and PCS. Regardless of the precise mechanisms, maladaptive coping styles such as rumination may increase perceptions of disability and reduce functional capacity, reinforcing higher PCS symptom reporting (Trahan et al., 2001). Indeed, this explanation aligns with research emphasizing the importance of coping processes in the progression of PCS (Jacobson, 1995) and, more broadly, the role of cognitive coping styles in an individual’s adjustment following illness (Soo & Sherman, 2015).

Contrary to expectations, reflection did not significantly predict lower PCS symptom reporting. On the surface, our results may suggest that self-reflection does not safeguard against post-concussive symptoms. However, a more plausible explanation may lie in the one-way predictive relationship between reflection and rumination. Specifically, reflection significantly predicts rumination, but not vice versa (Takano & Tanno, 2009; van Seggelen-Damen & van Dam, 2016). In other words, an individual who engages in reflection is likely to also ruminate about themselves, while those who ruminate do not necessarily reflect (Lengelle et al., 2016; Takano & Tanno, 2009; Şimşek, 2013). As a result of this unidirectional process, the benefits of reflection, such as enhanced self-knowledge, lowered distress and greater psychological adjustment (Jones et al., 2009; Trapnell & Campbell, 1999; Watkins, 2008), are negated by the negative effects of rumination (Takano & Tanno, 2009). Individuals become trapped in the well-established relationship between rumination and depression (Takano & Tanno, 2009), a pattern supported by prior research showing that when reflection transforms into rumination, it exacerbates low mood and diminishes the positive effects typically associated with reflective thinking (Nolen-Hoeksema, 1996; Van Seggelen - Damen & Van Dam, 2016). Indeed, it is possible that the inclusion of rumination prior to reflection in our regression model may have statistically absorbed the variance attributable to reflection, masking any potential protective effect. However, follow-up analyses in which reflection was entered into the model prior to rumination produced the same pattern of results, such that reflection remained non-significant, and rumination continued to significantly predict PCS. This finding suggests that the null result for reflection was not due to the order of variable entry, but may instead reflect a true lack of protective effect in this sample or a more complex interactive process that warrants further exploration.

Alternative explanations are also possible. Firstly, rather than exerting a direct effect on PCS, reflection may influence symptom perception through intermediary psychological factors such as resilience, self-efficacy, or cognitive flexibility (Jones, 2009; Watkins, 2008). While reflection may increase self-awareness, its benefits could depend on an individual’s ability to translate self-reflection into adaptive coping strategies. If cognitive flexibility (or some other factor) mediates this relationship, the direct association between reflection and PCS may not have been detectable in the present study. Secondly, although reflection is typically conceptualized as an adaptive cognitive style (Trapnell & Campbell, 1999), not all forms of self-focused thought are constructive. In some cases, individuals may engage in reflection that lacks resolution or clarity, leading to heightened self-focus or self-doubt. While this process is distinct from rumination in its intent and theoretical underpinnings, it may still resemble rumination in its psychological impact. For instance, unproductive reflection may fail to generate meaningful insight, thus undermining its potential protective effects. To this end, the use of the Rumination and Reflection Questionnaire (RRQ) may not have fully captured the nuances of reflection in the context of PCS. The scale does not differentiate between reflective self-awareness (adaptive) and excessive self-focus (potentially maladaptive). If the measure fails to distinguish between beneficial and distressing aspects of reflection, it may lack the sensitivity needed to detect its protective effects. Future research could employ alternative scales to better assess different types of reflection and their unique contributions to PCS symptom reporting.

The current findings pose important implications for researchers and clinicians alike. As rumination is transdiagnostic across various mental health conditions aside from depression (Nolen-Hoeksema & Watkins, 2011; Watkins & Roberts, 2020), it may explain elevated PCS symptomatology frequently reported by other psychiatric populations, such as PTSD (Lagarde et al., 2014; Stein & McAllister, 2009; Wilk et al., 2012), anxiety disorders (Thomas et al., 2022; Wallace et al., 2020), or somatization disorder (Donnell et al., 2012). Thus, future research should expand on the present results by clarifying the role of rumination in mediating the PCS symptom variability in different mental health disorders. To the extent that coping processes affect PCS development (Iverson et al., 2020; Jacobson, 1995), it is also critical to explore how other cognitive coping styles (e.g., catastrophizing) and their potential interactions may predict mTBI recovery outcomes. Clinically, our findings underscore the utility of evaluating an individual’s level of rumination following an mTBI to (a) assess their susceptibility to PCS and (b) identify rumination as a perpetuating factor in maintaining their PCS, which may be addressed by psychological interventions targeting maladaptive cognitive processes (Silver et al., 2013). Cognitive-behavioral therapy (CBT) is effective in reducing rumination by helping individuals recognize and challenge negative thought patterns (Querstret & Cropley, 2013; Watkins, 2016; Watkins & Roberts, 2020). Specifically for PCS, CBT can disrupt ruminative patterns by first validating the patient’s difficulties and their sense of self (Silver et al., 2013). The subsequent stage involves identifying cognitive dysfunctions and their contributing factors and ultimately enhancing an adaptive sense of self through learning healthy coping strategies and emotional regulation (Broshek et al., 2015; Silver et al., 2013). Indeed, it has been shown that an increased perception of self and greater self-efficacy are significantly protective against PCS (Kersivien et al., 2024).

Despite its unique contributions to PCS research, the current study has limitations that should be acknowledged to facilitate objective and accurate interpretation of the results. Firstly, our sample mainly consisted of females, undergraduate students, and young adults (M_age = 21 years). This will likely restrict the generalizability of our findings to other demographic groups, warranting the use of a more comprehensive study sample in future research (Edmed & Sullivan, 2012). Similarly, as people with a prior head injury were excluded, the present study used a non-clinical cohort only. It is unclear how our findings may be applied to mTBI individuals in clinical settings. The incorporation of a clinical comparison sample will thus inform differences compared to healthy controls, helping validate implications such as the effectiveness of reducing rumination to improve PCS outcomes in mTBI populations. Additionally, while reflection did not directly predict PCS reporting in this sample, it may still exert influence through indirect psychological pathways. Reflection has been associated with increased self-awareness, resilience, and adaptive problem-solving in prior studies (Trapnell & Campbell, 1999; Takano & Tanno, 2009). These factors may buffer against the negative effects of stress or somatic preoccupation, thereby reducing PCS reporting. Thus, even in the absence of a direct effect, reflection may influence PCS through mediators such as cognitive flexibility, self-efficacy, or trait resilience. Future research could examine these indirect pathways using mediation models to clarify the protective mechanisms through which reflection operates.

CONCLUSION

The results of the current study showed that higher levels of rumination predicted greater reporting of post-concussive symptoms above and beyond the impact of current symptoms of depression. This study highlights the role of poor coping styles in underscoring the relationship between depression and PCS in a non-clinical population. This supports the early assessment of individuals’ ruminative tendencies and the use of targeted interventions to address dysfunctional cognitive thought patterns. Indeed, future efforts identifying additional key processes linking pre-existing mental health conditions and PCS are a critical step toward improving outcomes for those with mTBI and the general public alike. These findings may inform clinical work with mTBI populations by highlighting the potential benefit of interventions that target rumination to reduce PCS burden.

FUNDING

None declared.

CONFLICT OF INTEREST

None declared.